GTM 157: How ClickUp Scaled 17x and Cut CAC 3× — GTM Systems, Building Incrementally and Going from PLG to SLG | Gaurav Agarwal

The GTM Podcast is available on any major directory, including:

• Apple Podcasts
• Spotify
• YouTube

Gaurav Agarwal is the COO of ClickUp, where he’s helped scale the company 17x in two years and cut CAC by 3x—all while leading across marketing, product, sales, and services. Formerly in banking and product roles, Gaurav brings a rare systems-oriented mindset to growth. He’s known for connecting the dots across GTM, scaling PLG engines, driving incrementality with rigor, and building cultures of experimentation, ownership, and speed.

Discussed in this episode

• Why growth is more about systems than marketing
• How ClickUp scaled 17x while reducing CAC by 3x
• The cultural shifts required to align GTM teams
• Incrementality testing vs. attribution modeling
• Building a true sales motion on top of PLG
• Input KPIs as a driver of experimentation speed
• How ClickUp is using AI across marketing and sales
• The concept of GTM as demand creation vs. harvesting

Episode Highlights

00:00 — Everyone is doing all the work—but you’re winning by chance, not by design

02:01 — Gaurav on seeing growth as a systems function, not just marketing

13:33 — Culture alone delivered roughly half of the 3× CAC reduction at ClickUp

15:53 — Gaurav defines incrementality testing vs. flawed attribution models

29:27 — Setting input KPIs so squads deliver a “win every week” keeps momentum high

32:53 — AI is being embedded across PLG + GTM workflows—from content briefs to deal signals

Recommended Books

• Shareholder Letters by Jeff Bezos

Referenced

• ClickUp: https://clickup.com/
• Gong: https://www.gong.io/
• Netflix: https://www.netflix.com/
• Robinhood: https://robinhood.com/

Guest Links (Gaurav Agarwal):

• LinkedIn:https://www.linkedin.com/in/gauravragarwal/

Host Speaker Links (Sophie Buonassisi):

• LinkedIn: https://www.linkedin.com/in/sophiebuonassisi/
• Newsletter: https://substack.com/@sophiebuonassisi

Where to find GTMnow (GTMfund’s media brand):

• Website: https://gtmnow.com/
• LinkedIn: https://www.linkedin.com/company/gtmnow/
• Twitter/X: https://x.com/GTMnow_
• YouTube: /@gtm_now
• The GTM Podcast (on all major directories): https://gtmnow.com/tag/podcast/

Thanks to our Sponsor: Qualified

Piper is the #1 AI SDR on G2.

AI SDR agents are changing the future of sales. And Piper is leading the way.

Piper is the top-ranked AI SDR on G2, trusted by hundreds of high-growth GTM teams — including many in the GTMfund community. She works around the clock to follow up with every lead, book meetings, and qualify prospects, all by reading real buying signals.

That means your team spends less time chasing and more time closing.

Piper isn’t just automating repetitive work — she’s transforming pipeline generation at scale.

Discover how AI SDRs like Piper can help you build pipeline faster at qualified.com.

The GTMnow Podcast
The GTMnow Podcast is a weekly podcast featuring interviews with the top 1% GTM executives, VCs, and founders. Conversations reveal the unshared details behind how they have grown companies, and the go-to-market strategies responsible for shaping that growth.

GTM 157 Episode Transcript

Sophie Buonassisi (01:57.966)

Gaurav, welcome to the podcast.

Thanks for having me, Sophie. Glad to be here.

Gaurav Agarwal (02:01.966)

Absolutely super happy to have you here and I got to jump off here with a big one because you helped click up scale 17x in just two years leading growth across operations including marketing product sales solutions and services and Most people still think of growth as a marketing function So when did you realize it was actually more of a systems problem? And what was that biggest shift that you made as a result?

Yeah, I think that’s a great question. I actually fell into growth by chance. I’ve always seen myself as a business guy. What do I love to do? I love to build systems that can generate customer value, which can then be captured to generate shareholder value. So I’ve always seen myself as a business guy. I left banking. My first job out of college was investment banking. I did that, but then I did product. Then I did customer development. Then I fell into growth and the way in growth was very

very, very young back in the day, like I’m talking 2013, 2014. So growth was just beginning to form. And the way we defined it was do whatever you need to do to win and grow the business. That’s how I got into the growth world. And it just happens that that was an e-commerce company. So it just happens that in an e-commerce hardware business, the levers at your disposal are more digital in nature, like marketing, acquisition, the funnel, the product, data, of that stuff.

But I’ve always seen myself as that. And even at ClickUp, I joined ClickUp as the chief growth officer primarily focused on the self-serve funnel and the self-serve machine. But today my remit has expanded. And that’s been the case for a couple of years now. And as the COO, I’m responsible for all things business. And then what are the levers at my disposal? They happen to be one on the digital side. So self-serve, funnel, acquisition, marketing, all of that stuff. But then it also happens to be one on the sales.

and the human touch side, which includes pre-sales, sales, post-sales services, all of that stuff. But to answer your question, I’ve always seen growth as a systems function. When I say systems, I don’t mean IT systems, but can you break down businesses into smaller set of systems and then connect the dots? So I’ve always seen growth as a function that connects the dots and owns business outcomes, and then uses all the levers at your disposal to generate outcome. And people who see it as marketing,

Sophie Buonassisi (04:25.602)

typically just only capture a sliver of what great operators can do.

Absolutely. yeah, the upside just continues to increase as you look and expand that across the go-to-market motion.

Exactly. That’s, mean, I think in some way, pretty much all businesses, like what does a great growth operator do or what does a great growth function do? You look at the data objectively, you place some bets and you iterate aggressively, right? And I think that’s, and you identify points of leverage, not everything needs to be optimized. And I think those are pretty much the fundamental tenets of running any business.

Okay, I’ve got a lot of areas I want to go down on that point, Gaurav, but take us back to the beginning before we deep dive on some areas. What did the actual go-to-market engine look like at ClickUp when you joined?

So when I had joined ClickUp, we were pretty much trying to do everything. We are still trying to do everything, but the main difference is at that point, these different bets, like we were trying to spin up a channel motion. So what has worked for ClickUp? ClickUp is a product-led growth company. We have a very strong self-serve growth funnel. Half of that is fueled through organic, the other half is fueled through paid media, advertising. Then, of course, you have sales teams. So when I joined, we were trying to do everything.

Sophie Buonassisi (05:48.408)

We are trying to build a channel motion. We are trying to build a success team. We had just raised money and we are trying to bring a of strong leaders in together.

The problem was though, even though we were bringing very strong personalities and very experienced leaders in, we were not connecting them together as a part of one singular motion. So everyone was trying to do their own thing on the side. And the problem with that is when you do that, you accrue bloat and inefficiency. Your payback goes up. Our paybacks were extremely bad, like closer to 60, 70 months back in the day. So since then I have brought it down to a place where

It’s very comparable to an efficient public company. It’s very efficient. They have brought down paybacks by multiple factors. And that has allowed us to build a great motion that we continue to build on top of. But what led us to that? I think we’re still doing the same things. We’re doing, in fact, more things now than what we were doing before. But the motion previously was a PLG and a sales motion, but everyone pulling in their own direction.

versus today the motion is still very much PLG and sales led. So motion is similar, but they’re all pulling in the same direction, which means a much better experience for your customer, a better use of resources, better messaging, all of that stuff. So motion has not quite changed. I mean, in many ways it has changed because pulling everyone in the same direction is not easy, but we have matured a lot in how we run the business.

When you say pulling in the same direction, how do you actually get people to row that boat in the same direction? What are you changing? What levers are you shifting?

Sophie Buonassisi (07:26.094)

I think a lot of it then comes down to like.

When a company grows really fast, like ClickUp was, when you are not matured, it’s a little bit like Wild Wild West. Everyone is doing whatever they can to work. The problem with that approach is you’re winning, but you can’t necessarily call out what is your exact motion under the hood. You’re not winning by design, you’re winning by chance. As we pull in the same direction, I think what becomes important is truly understanding, okay, where does each team play in the motion?

So what’s the role of PLG? What’s the role of sales dev? What’s the role of sellers? Where does post sales come in? Defining those guardrails that, you have to, we have to stick like services, for example, can play in this sandbox. So defining those guardrails clearly, setting the boundaries for what a team should be doing and what does great look like. Setting the boundary, but then also setting the expectation of what does great execution look like within that team.

And really, that’s the difference. You start pulling them in the same direction because then you define one motion where you take into account not one team and each team canvassing for their own codes, but you take into account the broader motion together. And then you stitch one customer journey that is comprehensive and that is consistent. And you know that which team operates at which part of the customer journey, which makes it super exciting for us because it’s fun. There’s no overlap. There’s no ugly. We used to call this

we call it eight layer cake. At some point, sometimes our customers would be touched by five or six people at the same time. And we’re like, why? So I think a lot of what pulling in same direction means you’re defining clear guardrails, expectations on what it looks like, and then strategically building the motion such that it delivers the best outcome for the customer and the business.

Gaurav Agarwal (09:22.626)

That’s fantastic. To get even more tactical, this is amazing, what did you do when you entered to actually layer that cake? How long does this exercise take? Who’s involved? What are you using to whiteboard if you’re whiteboarding?

I think the most important thing that I’m very grateful for is you can’t do any of these exercises if your leadership is not company first. And that’s one of the hardest things to do as we are all humans at the end, right? We all have ambitions and I have ambitions too. My leaders have ambitions too, but it’s okay to sometimes say that.

And we do this so well in our company. We say, okay, I, as an individual feel driven towards this decision. But if I take a step back and observe what’s happening as a third party, this is what we should be really doing. So the first thing you need to do to drive all of these changes, build that culture of intellectual honesty and objectivity where your leaders are able to acknowledge that we are all ultimately humans. But when we get together as a team, we are optimizing for the best decisions for the company. When you build a culture like that,

When your leadership works like that, there’s no one who is playing games or there’s no one who is trying to optimize for a silo. So really, really grateful for that. I have a very strong bench and I’m very proud of that. So that was the first thing. After that, if everyone believes in the same thing, then 50 % of the decisions become obvious. That no one is trying to canvas for a head count. No one is trying to fight for a book of the business. So then you look at it and you say, does it make sense? Like just use our common sense.

Does it make sense from a common sense perspective? 50 % of the noise you can cut by just operating out of common sense and not operating out of agendas. And I think that becomes very clear, especially if you put a bunch of smart people in the room. The remainder 50%, I would say, is very iterative and data driven. So you take a bet and you say, what happens if you move this team from here to here? OK, let’s try it out. Let’s try it out and see how it works out. So we place a lot of bets like that.

Sophie Buonassisi (11:28.834)

that are run on a quarterly cadence where we are constantly trying out new things, we are optimizing the motion, and then we go from there.

Brilliant. Brilliant. I love that experimentation. Often we think of experimentation more within the parameters of an existing funnel and existing department, but you’re almost moving the players on the chessboard and rearranging the organization as an experiment.

And the leaders, we all get into our room together and we decide that together. So like, hey, what do we do to make this motion efficient? Because at the end of the day, as a company that is quite late, we’re quite matured, ClickUp is well within the nine figures revenue, we continue to grow it. We continue to grow at 40%. So we’re doing great. But the chess board needs to expand. Each team needs to improve how they operate, which happens. But when me and my leaders, get together, we think about like, what do we move?

What chess pieces do we move between ourselves to see, can we make a better motion? Because the truth is, you have to meet a certain financial target. At some point, it’s not about optimizing for your own thing anymore, because the truth is, the size of the pie is limited. We have to distribute, there’s a fixed budget, and within this budget, we have to run the business and grow the business. So you have to play for the team. And that enables us to play chess.

between different teams versus just within teams.

Gaurav Agarwal (12:53.174)

Right. You’re ultimately one color. One team.

Yes, we all want to win together. I think that’s the thing that matters. It’s all the first team.

Absolutely. That’s where we cue all the sports analogies.

Yeah, not a big sports guy myself, but I know a lot of my peers are.

Yeah, well, let’s get them. So you’ve cut CAC by 3x. How did you diagnose what was broken, rebuild the system? Sounds like experimentation is a big part of that, but step by step, what does that process look like? Because right now, when everyone’s prioritizing efficiency, that’s such a huge lever that people want to pull on.

Sophie Buonassisi (13:33.71)

Yeah, I think it comes down to, I’ll repeat the answer I said, but then I’ll give more color to it. Once you get smart people in a room and you say play for the team, they start calling out the fat. how do you, everyone knows there’s fat within their teams. How do you build an environment and a culture where people volunteer that, there’s some fat sitting right here. Let’s sack, let me offer my fat because when we do that together, we are making the machine better. So

that culture of intellectual honesty and ownership goes a long way that gets you halfway there because you look at any GTM motion that’s bloated. I bet you that people have had people are sandbagging. They’re not thinking through. They want more. They want more headcount because it feels comfortable. But if you set the culture as one where we are going to raise the bar, especially in the world of AI, if you’re not raising the bar and demanding more, you’re going to lose out eventually anyway. So then you’re just delaying death.

So you are better off saying, we have to raise the bar, we have to be intellectually honest, we have to play for the team. If this is how we operate and this is how we hold each other accountable, automatically waste begins to emerge. That gets you halfway there. Honestly, was out of the 3X reduction we did in GAC, half of it was just people accepting that we have to raise the bar and we have to do things differently. So that’s half of it. The other half of it is very interesting.

and I talk about this quite a bit. A lot of operators are playing the attribution game, which sucks because you are trying to play the credit game. But here’s the other thing, attribution is one of those memes where you might have seen it, like the bottom 5%, the top 5%, and the medium is just trying to play with attribution to play blame games. I actually play attribution game too. But if done well,

It can be one of the most strongest levers to power growth. And the way we look at things in ClickUp is you need to understand what part of your portfolio is truly driving results versus what is not. And we use the word incrementality. Let me give you an example. Let’s say there’s display ads is a good one. This happened back in the growth marketing days. Display ads would show up on every website you go to.

Sophie Buonassisi (15:53.346)

They’ll just retarget the hell out of you. There was a time when if you go to any website, those banner ads would follow you. So the attribution game would say banner ads drove the purchase because they were this person, Sophie saw these banner ads 20 times. That’s the fool’s way of doing attribution. The smarter way doing attribution is to run a test. We call it incrementality test where I’ll show Sophie banner ads, but then Gaurav will not see any banner ads. And then let’s see.

yeah.

Sophie Buonassisi (16:22.434)

Did Sophie buy more compared to a control which is Gaurav? And what you’ll most likely find is no banner ads were just showing up because they knew how to show up in front of Sophie. But that does not mean that drove the intent or that build the desire for Sophie to purchase. understanding, I have a complex portfolio that spans across marketing channels, self-serve marketing, B2B marketing. There are different

then there are human resources that are being deployed everywhere, sales, et cetera. But you need to understand what’s the incrementality of each intervention. What part of my portfolio is driving returns? What part of my portfolio is not driving returns? When you operate your GTM machine like a portfolio, think like a financial services portfolio, where like you know that, okay, these are the bets I’ve made, and these are the bets that are driving return, and return not measured as attribution, but return measured as incrementality. That’s how we got

got to the second half of the 1.5x cap reduction happened because the culture is great, raising the bar, people are honest. The other 1.5x, we got it through a ton of incrementality testing.

That’s, mean, first incredible that you got 1.5 X just by simply changing the culture. That speaks volumes to what that alignment does and how that could benefit other organizations too. The incrementality is fascinating also. I love the analogy around more of a financial analysis and portfolio in general. think that simplifies it. The beauty of simplicity is it brings it down, it boils it down to that level.

It does exactly. It makes it very simple in the sense that you need to know what part of your portfolio is driving results.

Gaurav Agarwal (18:10.762)

Mm-hmm. And when you ran that incrementality testing, what were the biggest levers or learnings involved?

So that’s a good one, right? will not everything. So as a growth guy, I get this question a lot. Yeah. Are you data? Everyone should think of me or other operators that talk about data as people wanting a very accurate, very precise data driven thing. I actually think that’s a myth. Sometimes even just having a qualitative data set is great. Like go talk to customers and say, did you like the sales experience or not?

And yes, no, yes, no, yes, no. And then you’ll be like, yeah, this sales experience that we are delivering isn’t the great sales experience. So it’s more important to be directionally accurate with data than fully data driven. You have to be data inspired, if you may. Can you look at the data, try to piece the story together? And a 70%, 80 % accurate story is great. One of the best things that I like to read, these are shareholder led.

shareholders’ letters written by Jeff Bezos. And in one of the things he talks about, can you make decisions at 70 % resolution versus trying to shoot for 100 % accuracy and precision in everything that you do? So to answer your question around incrementality testing, everything looks different. For example, testing for advertising is very different than testing for, say, sales dev versus testing for a sales intervention. And you can spend countless hours

I know people, I know people who are extremely smart, PhD, they spend hours and hours and weeks and weeks designing the perfect test. Worse is sometimes you say, you know what, let’s just go run an imperfect test because we will learn something in the process. The outcome isn’t decisions. People think the outcome of testing is decision. The outcome of testing is learning. And whatever allows you to learn fast, you should be moving in that direction because when you learn something, there are so many tests we run here.

Sophie Buonassisi (20:13.634)

where we stop the tests midway because we learned something that we did not know, then we iterate on it and we ship another one. Because we are constantly, what you want to build is an organization that is constantly learning. And learning at the top of the funnel is very different than learning how success works, or learning how success is very different than learning how sales works. So you have to design each of these tests and sit with the ops teams, sit with the leadership.

But I guess what I want to say is even using the word design, the test feels like the wrong word. It’s more around sit with them and try to understand what’s happening and what can we not explain? Okay, great. Now, how do we understand that? We can run a test, we can talk to customers, we can look into the data, but it’s constantly trying to learn around the corners where light bends.

Super interesting. And I love the emphasis on qualitative because think companies often assume as you mature and grow, you lean even heavier into the quantitative side. And that’s not wrong. But you just stress the importance of qualitative even at more of that scale up stage, which is really interesting. Also an interval that’s lower, know, a stat sig typical level would be like 95%.

but that 70 % allows you to probably launch quicker too.

Well, I think if you have the luxury of running off, like there are lots of tests that we run that are at 95 % as well. And we see that through, but I also see a lot of people not test because they don’t have enough data. And I’m like, you can just, you can just rank this in on a five point scale, just do that. Like, and then see what’s working. So it’s about taking an analytical mind and trying to find a pattern. It’s all building a business is about building a great system.

Gaurav Agarwal (21:50.199)

Yeah.

Sophie Buonassisi (22:07.948)

And if you’re building a system, need to understand the patterns that govern that system. And you could argue that I don’t have stat-sig data and I don’t have a million data points and I can’t tell you what the pattern is. But the truth is pattern exists in everything. And you may not know the exactness of that pattern, that’s fine. But you should directionally know what the patterns are. Like I don’t need a ton of data to tell me that when my cat is hungry, she wobbles her head.

That’s just a pattern and she has done that like ten times and I know that that’s what she does when she’s hungry. So I think it’s it’s building that muscle where you’re constantly seeking learning and knowledge and you’re trying to find the pattern. If you can do it with high degree of precision and data do it. But if you can’t then of course you need you need to do the best you need to make the best with what you have.

Pattern recognition, growth is pattern recognition. I remember, I used to actually fall on the very quant heavy side of the camp where I was running experiments at scale and measuring qualitative, but heavily skewing quantitative. And when I made a shift to measuring qualitative, I remember just having such a, such a challenge mentally overcoming it. And I chatted with one of my friends as the CMO at the time of a series eight company. Now they’ve gone on to raise and she said, Sophie.

pattern recognition.

That’s really well said.

Gaurav Agarwal (23:29.804)

We have a slide in our deck for the board that is the vibe pattern slide. And it’s where we just collect our qualitative feedback that is pattern recognition, exactly like you said. And so now I do that too. It’s the qualitative pattern recognition. That’s fascinating.

Yeah, I mean, it’s what you say, like first, when you’re young, you think data is everything. And then when you grow up, you realize that finding patterns is the most important thing within the data you have, but that’s really well put.

Absolutely. Pattern recognition. And now take me through more of the motion. Cause you were, you mentioned you were very PLG. You built up the SLG. That is a very common transition that a lot of companies and startups, even scale ups want to make. that transition from PLG to sales assist to sales led. Take me through what that looked like for your transition.

Yeah, so even you scale your PLG funnel, right? Your PLG essentially becomes your top of the funnel. So people are signing up for your product. They use your product. And that’s how your top of the funnel gets defined. And then sales is job is to then attach itself to that revenue machine and maximize LTV. So people come in, they sign up, but then you need to fit some of them will raise their hand hand raisers as we call it. They’ll go straight to sales and then.

a seller would touch that, we’ll get about a 10x lift in LTV. So that’s the big reason to overlay a sales motion on a PLG motion, because there are customers who want to talk to someone, especially if they want to buy a bigger deal. They want to buy an enterprise-wide license, so you need a seller. But in that case, a seller is basically taking a hand raiser. Then the next motion that you build on top of that is what you might call product-led sales, where you’re constantly looking at

Sophie Buonassisi (25:23.746)

triggers within the product, like what are people doing within the product? And instead of waiting for them to reach out to you, you are constantly reaching out to them and generating demand. And that’s what I would call like true sales in a PLG motion. It’s still, and it’s sales led because you’re creating growth out of product signals. The next phase that we are in the active phase of nailing, and I think that’s what a lot of PLG companies never nail, is how do you then move

to a true sales outbound motion, where sales is not just teethered with the product data, but the seller can identify what the white space in the account is, who are the ICPs, and come over the top and truly make top-down sell. That’s the motion we’re trying to nail right now. So as of today, a lot of our motion is still very much PLG heavy and very product-led sales heavy. We land in a department, we know how to go big in that department. We are very user-focused.

But now we building that lens that is very buyer focused, where our sellers can go sell a top-down deal as they would have done in a non-PLG funnel.

Right, that makes sense. Were there particular signals when you got started with that sales assist motion that you found were check marks and you would now, I know you advise a lot of companies, advise companies to look at those specific signals first when building out a sales assist emotion?

most important signals that we have found that a sales team should be operationalizing is really the growth within the workspace or the account. And I’ve heard this validated across so many times that when a workspace, if it’s a collaborative product, if the usage is going up disproportionately, or if it’s inviting more and more people disproportionately, then you know that that’s the time to attack a workspace because the workspace is experiencing growth and they will be open.

Sophie Buonassisi (27:22.584)

to you helping them out. That’s one. Second, it’s always better to reach out to them from a perspective of, can help you achieve more out of the product versus trying to sell. And if you take that second angle, then in the world of AI where outbound can be programmatic, you actually want to reach out to them as many times as possible through meaningful advice. That I saw you were trying to do this in the product. I believe we can help you do this this way as we have this customer drop in a case study.

In fact, this will be a very interesting learning.

We learned this is quite intuitive now in a world where cost of sending emails isn’t very high.

you shouldn’t be trying to rotate too much on the signals. What I mean by that is test every damn signal that you can. So many teams spend so much time scoring every signal and running a data science model and running and so what instead of all of that, why are you trying to pick the two signals out of 10? Just run a quick experiment on all the 10 and see which one of these signals or which one of these experiments resulted into pipeline.

And then they picked the one that resulted into pipelines and the ones that did not keep into pipe result into pipeline. They were not the right signals anyway. So try and be more action oriented, ship more, test more, were sustained that data science land trying to, trying to predict each signal because your cost of sending an email, the cost of outreach is literally zero. And with AI, we have built a ton of AI SDRs. have built a ton of, no, SDR would be the wrong word. We built a ton of AI outreach.

Sophie Buonassisi (29:02.774)

systems that allow us to personalize every signal, every account really well. So we are shifting heavy towards analyze less, throw more stuff out there and see what can you learn from it.

I love it. Culture of experimentation. How do you actually instill that culture of experimentation or speed in people too?

Sophie Buonassisi (29:27.726)

I think some of it is you can’t quote someone on speed if they’re not wired that way. There are some people who just need more structure and I think that’s okay. But then assuming that you do have the right people, the way at least I try to drive that culture of speed is through KPIs. My teams, a lot of my teams don’t just have like you have to meet this end goal.

They have activity as a core KPI as well. So for example, just today I was in a growth product meeting and I said, I need a win every week. So there are four teams in that. were four squads and I said, each of the squads need to deliver one win per week. So then what do you do? Then you calculate that if I take five shots at a 20 % success rate, I’ll have a win every week. So what God was really telling us is you need to take five shots at the goal every week, or you have to get really good at win at improving your win rate. That’s not going to happen.

So what Gaurav is indirectly holding us accountable to is can we make five bets every week and what do we need to make that happen? So I think we hold ourselves accountable to that culture of speed and velocity by holding ourselves accountable to activity KPIs or input KPIs. And then of course, if your input KPIs are great, your output will go up. But there’s no point trying to hold accountability towards an output KPI and not inspect the input KPI because

garbage in, garbage out. So we, I personally obsessed a lot more about the input KPIs because if you put in the hard work, if you put in the right, if you’re inspecting the right things that put in the right effort, outcome will follow.

Gaurav Agarwal (31:12.174)

And that’s actually a similar model now with AI. It’s all about input, which controls output.

It’s all about input. It’s all about fine tuning AI to your use case, which is all about inputs. What inputs do you give AI? What context do you share with AI? And that’s what would decide who ends up winning because everyone has access to the same models. But the ones who can give AI really relevant context are the ones who are going to win.

How are you using AI in ClickUp? Any specific areas that you found a lot of leverage or incrementality around?

Great, great, great point there. I love the word incrementality that you just used. So we use AI pretty heavily and we use AI across our entire PLGTM function, PLG plus GTM function as I call it. We even use it on our EPD side of the house. I think a lot of that is because we work in ClickUp. ClickUp at the end of the day is a work management platform where you can orchestrate work and workflows. And if you work within ClickUp,

Internally, we haven’t rolled this out to our customers yet, but we’re in the process of rolling it out. AI can intervene at different parts of your workflow and help you do work faster. So our teams are writing hundreds and thousands of blog briefs, SEO briefs on ClickUp directly. We are using, of course, we are also using state of the art AI tools. So like we’re using tools like Kong, et cetera, for deal inspection. But then we also bring some of that data into ClickUp.

Sophie Buonassisi (32:53.42)

and then click up all our AI internally consumes that information and tells us that, these are the deals you should be paying attention to. So we are trying to automate as many workflows as we can. They happen to be deeper on the marketing side, on the content generation side, on the brief generation side, but we’re trying to now go into deeper use cases, like solving AI for analytics is one of the most important things on my mind right now, because then you can connect analysis to generation.

and you can automate that entire workflow. But we use AI pretty heavily across everything. Our sellers use AI for generating first draft emails, account research is all AI driven. So what else do we do? Signals, with signals, with signals should we go after with what messaging? All of that is then using AI. We are able to customize and personalize. ClickUp has like five different ICPs, but 60, 70 use cases.

We are a very horizontal platform, so we use AI heavily to personalize messaging across each of those use cases.

Right, right, that’s fantastic. Brilliant.

Brilliant. Well, I’m sure we’ll be seeing increase too as we continue to gain incredible leverage from AI.

Sophie Buonassisi (34:13.004)

beginning, which is beginning. I think we haven’t even scratched the surface because I say that because what I’m tracking right now is how many of my employees are AI native. And that’s going to become extremely, extremely important because look at the people entering the workforce right now. They use AI the same way as I remember like people using internet.

and there were lots of people who even struggled with computers at that point. So one of the most important battle cries we have is, our workflows are getting automated, et cetera, but everyone needs to find a way to 10x themselves using AI. And you don’t do it just because ClickUp is asking you to do it, but you have to do it if you want to stay relevant in the future. And it’s not that AI will take your job or something, but you are so smart. What if you can have AI go do 80 % of the work that is repetitive? So you can then focus on

managing all the different agents and truly maximizing your own potential. So we haven’t even scratched the surface on that. And that’s one of the huge battle cries for us is how do we take our teams and make them truly AI native?

Mm-hmm. Mm-hmm. Yeah. I love the saying. hear this, I mean, I hear it many times, but it’s that AI isn’t replacing yourself. It’s replacing people that aren’t using AI. Exactly. That’s a big difference. How are you incentivizing people to actually adopt AI? And then more importantly, and what I’m really interested in, and everybody listening is interested in, we’ve been swapping a lot of notes offline off the podcast about it is how are you…

sharing those learnings of everybody’s individual experimentations with AI.

Sophie Buonassisi (35:59.586)

We are very public about everyone using AI. So we have chat groups where people are talking about what are they trying. People ask for help. So I think there’s a community led motion going on where people within the company are just teaching each other how to use AI. And I think that’s the most organic way to go because eventually sooner or later, it’s like typewriters to computers. Everyone will have to learn how to use computers or desktop to laptops. Everyone will get on the internet. So whenever a wave like that happens, people get used to it.

Just that the ones who get used to it faster are the ones who capture a lot more value than the ones who stay behind and take forever to get there. And we’re putting in some incentive structures in place, et cetera. There’s coaching, there’s learning, all those things. But we also understand that it’s like going to the gym. There’s no amount of incentive you can give me that will say, go do this. If you really want to do this, you will do this. Otherwise you won’t. And what we are focusing on is how do we take the five to 10 % people who are actively building?

and put them in the center so their magic dust can rub off against other people and they start building as well. But I think that’s the way to go. It’s osmosis. When everyone around you starts doing it, you will start doing it too. And that’s just how these waves go. We are applying a lot of top-down mandate and messaging. For example, every team needs to automate their workflows using AI. Every team needs to come up with a hack per month that they have used AI for.

So we are applying those top-down pressure, but ultimately what will drive true usage is going to be osmosis, where people who have found success will talk about it and get the word across.

That’s where I feel like the speed at which you operate is a huge benefit to you because you are creating that osmosis quicker than a lot of other organizations. And as such, you are adopting AI quicker, which is giving you greater incrementality and so forth. So it all trickles down from your culture that you’ve built.

Sophie Buonassisi (38:01.038)

Yeah, the culture of speed and culture of ownership and accepting that we have to raise the bar. I think that’s one of the most important things. Like if you don’t raise the bar, you’re doing a disservice to your people.

That’s a great way of putting it. And you’re building some incredible culture, but also incredible systems at ClickUp. What’s one system that you’ve built that other SaaS companies should copy?

Sophie Buonassisi (38:31.754)

It’s a great question. I’ll contradict myself that I said before, but I think this is important. That I spoke a lot about data, right? If you go to a growth machine, if you go to Robinhood, if you go to any growth machine, that entire GTM is run out of a very, it’s almost like sitting on the trading floor. I used to do that. So I know what it feels like when you’re trading bonds, when you’re trading complex instruments and you have signals popping everywhere and this is what you do and this is what you do next.

And that’s what you need to run a highly effective business where you can beat competition. So it’s done really well in trading. Then some, the best of the best operators know how to build such data-driven systems within consumer growth. Like Netflix knows what show to recommend to you, Sophie, that you would have a higher chance to watch it, right? We are trying to that across the entire stack, all the way from what’s the right ad we show, is this the right person to show an ad?

does a great job.

Sophie Buonassisi (39:30.946)

When they come in, what’s their predicted LTV? Should we be spending more time with them? What’s the next best intervention? And then what’s the right time for a seller to reach out to them? And when the seller reaches out, what should that messaging be? And how should that account be distributed once that first sale has happened? Who should the account go to? And what do we expect from that account? And within the account, who should we be speaking to? And is the account trending healthy or not? And should customer service and success be intervening now or waiting for

the account is healthy. So should we be playing a proactive engagement motion? Or if the account is not healthy, should we be putting on technical resources on it? So we are trying to build this end-to-end data-driven map for our customers across each point in their lifecycle that allows us to deploy our resources really well and ensure that we are always delivering a great customer experience across each step in the journey.

I love it. kind of sounds like a customer experience constellation in a way where it’s not linear.

We actually, the code name for this is constellation. Because we understand that our customers are actually constellations. then we have to, we have to like, they’re all at different points in their journey. And we’ll have to ensure that each star gets the most love they can. So the internal project for this is we call it private constellation, but it’s because it’s about mapping the universe of your customers and then knowing what to do with each node of that constellation.

Yeah

Gaurav Agarwal (41:00.812)

Very cool. love that. fantastic. First of all, insight around one recommendation. I feel obliged in a way. I really want to double click on that and ask you a couple more pattern recognition questions, which seems to be a superpower. And you’ve seen inside other fast growing SaaS companies, in addition to ClickUp, what are a couple patterns that you’ve seen on repeat that more founders should be copying or taking inspiration from, maybe not copying?

You mean from a business perspective?

From a business perspective, from a metrics perspective, from a go-to-market-motion perspective.

Sophie Buonassisi (41:44.59)

I’ll give you one that’s on the people side, but then I’ll also think of something on the business side that has stayed consistent with me. One hard lesson that I have learned that I wish I could have learned slightly easier, but I keep relearning that over and over again, you need, there’s something that I use, I call it startup physics. A company grows, a company that grows 40, 50 % every year, every two years, the company would double.

That means the scale of operations, the challenges, the business goals, everything would double. Most humans don’t have the capacity to double their skill set and double their capacity to, most humans don’t have that growth rate. So you have to push your teams incredibly hard, even if they don’t like it, if you want them to be along on this journey with you. So one hard lesson that I’ve learned is you have to grow your people much more than you think you have to.

and otherwise you lose some people behind and you will outgrow your people. And that’s why a lot of companies, they churn out people every two years. That happens all the time at ClickUp. If people don’t work out with us, they’re not here after six months, but the ones are average tenure among employees is three plus years, four years. But that means we have to incredibly push hard on those people that are here because otherwise they’ll be left behind.

And there’s a reason why everyone just churns and burns in a startup machine. Two years is the average lifetime of an employee. So one lesson that I’ve learned the hard way, the metric that you have to pay, are you outgrowing your people? And if that’s happening, then be honest about it and attack it. Now on the business side.

I think this will sound very intuitive, but there’s only so much you can get from your existing customer base. So we talked about PLG, SLG, all of that stuff. I actually look at it slightly differently. I looked at it as demand creation and demand harvesting. You can have an SLG funnel where sales is creating demand, but then harvesting might be happening through self-serve or you can have a self-serve funnel like a classic lead gen.

Sophie Buonassisi (43:59.874)

where leads and demands are generated through marketing and sales is harvesting like a primarily inbound funnel. Or you can have a sales led only motion where sales is generating demand and also harvesting. So there’s all combinations of stuff possible, but I always look at demand gen, like what is generating demand? What is creating demand and what is harvesting demand? And the mistake I’ve seen a lot of people make is they think they’re both equal, but they’re not. Creating demand is significantly harder.

among people who don’t know about you versus harvesting demand that already exists about you. When I say harvesting demand, I don’t just mean it for the category. I mean harvesting demand for your product. So creating demand, creating awareness for your product is significantly is to my directional pattern that I’ve identified. Creating demand is about four to five times harder than harvesting demand.

where the person is saying that they already want your product and you’re just helping them close. And wherever possible, don’t think of your funnel as a PLG or SLG or a marketing and sale. Sales break it down into what within sales is creating demand and what within sales is harvesting demand or closing demand, what within marketing is creating demand and what within our growth is harvesting demand. And then look at your budget allocation. And you might find

that you are treating them the same and that’s wrong. The group that is closing demand needs to be significantly more effective and efficient than the group that is creating demand. So that’s a KPI that I don’t see a lot of operators or founders watch, but it’s extremely important that when you look at your machine, revenue machine, you break it down into creating demand versus harvesting demand and watch those ROI data really clearly.

Would you break that down into activities, if you will? So for example, a specific activity within marketing and then a specific activity within sales. Just thinking what if there’s overlap between the departments? So one department is harvesting in one area, but it’s also creating demand in another. Are you actually tagging them accordingly and then shifting budget to that specific initiative?

Sophie Buonassisi (46:22.254)

Exactly. would, I would be tagging not just individual activity, but a group of activities and then tag. Is your a hundred percent right? That’s exactly what’s happening under the hood. There are sellers who are creating demand. There are sellers who are harvesting. There’s marketing that is doing true prospecting. There’s marketing that’s just doing retargeting like the banner ads. You can’t say the banner ad is as effective as a YouTube ad that planted the seed in the first place.

is the same thing. Like you can’t say a person receiving an order, a transactional seller is doing the work that is as hard as a person who is trying to talk to a champion and get access to a buyer and do a bigger sell and creating demand in that process. understanding what is creating demand and understanding what is harvesting demand is extremely important. And you’re right, you’ll have to actually tag it at an activity level.

and then use that activity to figure out like, okay, what’s my ROI on this activity and what’s my ROI on this activity? And typically if I expect, I’ll expect like a 4X difference in that ROI. If closing the demand is say an ROI of 1X, sorry, if creating the demand is ROI of 1X, I would expect an ROI of 4X on closing the demand.

Gotcha. Okay. And where do feel like…

That’s all anecdotal based on my pattern recognition through my career.

Gaurav Agarwal (47:50.594)

No, it’s fantastic. mean, you’re getting even more granular than the typical here’s a budget for marketing, here’s a budget for sales allocated accordingly. You’re really looking at what’s actually driving revenue, irregardless of department. Where do you feel like the biggest drop is? So for a startup, for example, when they’re thinking about budget allocation between marketing and sales, what would you advise them to do rather than just, you know, drop a bag on either side of the fence?

I’ll just ask them, look at your motion carefully and look at what’s creating demand and what’s harvesting demand. And then based on that, make the allocation. Actually, there’s a third layer in SaaS. You could say, create demand, harvest demand, and retain customers. And you should be paying less for retaining customers. You will, right? There’s a reason why sellers get paid. Sellers are the heroes of modern-day SaaS, right? Because they are the ones who generate demand and close demand.

You want a significantly higher ROI on retaining customers. Then you want lesser ROI on closing demand, and then you want much lesser ROI on creating demand.

Super helpful. Gaurav, this has been incredibly tactical, data-driven, incrementality. There’s a lot of just notes I’ve got overall from you and that we really appreciate the time overall. Last question from my end is, you know, I’ve heard you talk, I it was on LinkedIn actually, you shared that the highest value activities for your time is something that you analyze. Yes. You also analyze patterns around, so you’re looking to shift.

lower leverage tasks off your plate. What is your framework for doing that?

Sophie Buonassisi (49:36.79)

I think you’ll have to realize that one thing that I see a lot of people make mistake around is you have to do everything yourself because you want the highest quality outcome in everything. But if you take a true systems mindset, what you’ll find is there are points of leverage. Not everything that just because you can attack and fix everything doesn’t mean you should be the one attacking and fixing everything. You should be the one attacking and fixing things that have the highest ROI for that.

per unit time spent, right? So one part of scaling yourself is also constantly finding ways where I have to constantly look at what are the highest leverage activities that I should be doing. But quality has to be high class no matter what. But of course, like I have constraints on my time and energy. Energy ends up becoming a bigger constraint than time for executives because there’s only so much mental energy you have left before your quality of decision making starts coming down and your thought process comes down.

So what you’re really constrained by is not time. Time is of course something you’re constrained by, but you’re also constrained by how many great decisions can you make every day. That means your objective function is not trying to do everything, but ensuring that the most amount of stuff gets done. I will tackle the high leverage stuff, but then I also need to pass on my Lego blocks to other people who have the hunger, the ambition, the quality, and let them make some mistakes if they may. But my job is to coach them and get them,

get them to be better. But the goal isn’t to like do everything by myself. The goal is to ensure that everything gets done with highest quality on the stuff that matters. And the stuff that is perhaps low leverage, let others learn and grow from that. So that has been extremely helpful for me as I’ve tried to scale myself.

I can imagine that would be helpful also with the objective of leveling up your employees and helping with that retention overall.

Sophie Buonassisi (51:35.906)

You need to give them a playground and then you need to be there as a coach, not just as someone who is judging their activity, but hey, let me tell you how to do it. And that’s how they get better. And as they get better, you can pass more to them. Everything is connected and that’s how you stretch them. That’s how you raise the bar that allows you to scale yourself and take on the next demons that you need to slay. But what you’re doing is you’re constantly creating this ground where everyone is learning and growing together.

Incredible. Well, Gaurav, this has been a fantastic conversation. Really appreciate the time. Where can people find you if they want to follow along or get in touch?

I’m on LinkedIn for a search, my first name, last name and ClickUp and I think you will most likely find me. But please do connect. You can also email me directly, gauravmyfirstnameatclickup.com. But Sophie, thank you so much for having me. This was the pleasure and love the fact that you went so deep into some of these tactical things. Thank you everyone.

Have a great one. We’ll see you next time.

You too, Sophie.